This application addresses NIH RFA-OD-09-004: NHLBI Participation in Research and Research Infrastructure "Grand Opportunities" (RC2) and specifically responds to the NHLBI RC2 topic of Comparative Effectiveness among adults with atrial fibrillation. Atrial fibrillation is the most common clinically significant arrhythmia in adults and one of the most potent risk factors for ischemic stroke and other systemic thromboembolism. It affects 2.5-5 million Americans currently and its prevalence will increase markedly with the aging of the population nationally. Numerous randomized controlled trials in selected populations have compared stroke prevention strategies (e.g., vitamin K antagonists, antiplatelet agents, other) and pharmacologic arrhythmia control, and numerous earlier phase studies have assessed emerging non-pharmacologic interventions (e.g., catheter ablation and surgical ablation). However, controversy remains about how the results apply to the more diverse and sicker population of atrial fibrillation patients treated in typical clinical practice. Furthermore, existing randomized trials have not been able to examine the variety of different combined treatment approaches or provide realistic estimates of potential risks and benefits in representative populations in the contemporary treatment era and relevant patient subgroups. In addition, we lack robust risk stratification schemes that have been validated among large, diverse populations and that address the significant variability in the risks of stroke and treatment-related outcomes among individual patients. Finally, we have limited data about whether health disparities exist within clinical populations of atrial fibrillation, both with regard to treatment strategies and the influence of selected patient characteristics on atrial fibrillation-related outcomes. To address these major knowledge gaps highlighted in NIH RFA-OD-09-004, we will leverage our expertise from the Anticoagulation and Risk Factors In Atrial Fibrillation (ATRIA) Study, the National Heart, Lung and Blood Institute-sponsored Cardiovascular Research Network (CVRN), a multicenter, health plan-based research consortium, as well as data from sentinel randomized clinical trials in atrial fibrillation to create an unparalleled multidisciplinary team and resources. The proposed project will establish a unique collaboration among population-based researchers, trialists, and methodologists as well as a unique research platform through a carefully characterized contemporary cohort of ~40,000 adults with atrial fibrillation to improve comparative effectiveness studies, risk stratification of outcomes, and the planning and conduct of future observational studies and randomized trials for atrial fibrillation. We propose to accomplish the following three Specific Aims: Aim 1: Develop and test novel risk stratification schemes for thromboembolism (ischemic stroke and systemic thromboembolism) in patients with atrial fibrillation from a large-scale community-based cohort and additionally validate the resulting risk models against rich datasets from randomized clinical trials in atrial fibrillation. Aim 2: Establish and characterize a contemporary registry of incident atrial fibrillation within very large, diverse community-based populations to provide critical insights into current outcome event rates and practice patterns, potential health disparities, and to facilitate more rapid enrollment into future effectiveness studies and clinical trials as well as development of improved risk stratification models. Aim 3: Identify and validate optimal design and analytic approaches to reduce confounding and bias for comparative effectiveness analyses (e.g., different antithrombotic strategies) in observational studies of patients with atrial fibrillation. PUBLIC HEALTH RELEVANCE: Atrial fibrillation is the most common clinically significant arrhythmia in adults and one of the most potent risk factors for ischemic stroke and other systemic throm-boembolism. It currently affects millions of Americans and its prevalence is expected to increase substantially during the next several decades. While multiple strategies have been developed to reduce the risks of atrial fibrillation-related adverse events, existing risk stratification schemes are of limited clinical utility in guiding treatment decisions. Furthermore, few data exist about the comparative effectiveness of different therapies in clinical practice. The proposed study will provide a critical set of research resources to improve risk prediction, facilitate more valid comparative effectiveness analyses, and streamline the planning and conduct of future randomized clinical trials in atrial fibrillation.